SoftClipsToSplitReads error: mem: invalid option -- 'K'

[kostasgalexiou]

Hi,

I am running GRIDSS through another tool and I am getting an error when running the CallVariants java file, at the step of executing SoftClipsToSplitReads

Part of the log file showing the error (here you have the full log file: .nextflow.log):

  [Fri Aug 23 09:38:25 CEST 2024] Executing as kalexiou@kalexiou-IRTA on Linux 6.5.0-44-generic amd64; OpenJDK 64-Bit Server VM 22.0.1-internal-adhoc.conda.src; Deflater: Intel; Inflater: Intel; Provider GCS is not available; Picard version: 2.13.2-gridss
  [Fri Aug 23 09:38:25 CEST 2024] gridss.ComputeSamTags done. Elapsed time: 0.00 minutes.
  Runtime.totalMemory()=285212672
  INFO	2024-08-23 09:38:25	SAMEvidenceSource	Identifying split reads for test.bam
  INFO	2024-08-23 09:38:25	SoftClipsToSplitReads	
  
  ********** NOTE: Picard's command line syntax is changing.
  **********
  ********** For more information, please see:
  ********** https://github.com/broadinstitute/picard/wiki/Command-Line-Syntax-Transition-For-Users-(Pre-Transition)
  **********
  ********** The command line looks like this in the new syntax:
  **********
  **********    SoftClipsToSplitReads -WORKER_THREADS 4 -INPUT ./test.bam.gridss.working/gridss.tmp.tagged.test.bam.sv.bam -OUTPUT ./test.bam.gridss.working/gridss.tmp.splitreads.test.bam.sv.bam -REALIGN_EXISTING_SPLIT_READS false -COMPRESSION_LEVEL 5 -CREATE_INDEX true -CREATE_MD5_FILE false -GA4GH_CLIENT_SECRETS client_secrets.json -MAX_RECORDS_IN_RAM 500000 -QUIET false -REFERENCE_SEQUENCE genome.fa -VALIDATION_STRINGENCY STRICT -VERBOSITY INFO
  **********
  
  
  [Fri Aug 23 09:38:25 CEST 2024] SoftClipsToSplitReads INPUT=./test.bam.gridss.working/gridss.tmp.tagged.test.bam.sv.bam OUTPUT=./test.bam.gridss.working/gridss.tmp.splitreads.test.bam.sv.bam REALIGN_EXISTING_SPLIT_READS=false WORKER_THREADS=4 VERBOSITY=INFO QUIET=false VALIDATION_STRINGENCY=STRICT COMPRESSION_LEVEL=5 MAX_RECORDS_IN_RAM=500000 CREATE_INDEX=true CREATE_MD5_FILE=false REFERENCE_SEQUENCE=genome.fa GA4GH_CLIENT_SECRETS=client_secrets.json    MIN_CLIP_LENGTH=20 MIN_CLIP_QUAL=5.0 PROCESS_SECONDARY_ALIGNMENTS=false REALIGN_ENTIRE_READ=false READJUST_PRIMARY_ALIGNMENT_POSITION=false WRITE_OA=true ALIGNER=EXTERNAL ALIGNER_BATCH_SIZE=500000 ALIGNER_STREAMING=false ALIGNER_COMMAND_LINE=[bwa, mem, -K, 10000000, -L, 0,0, -t, %3$d, %2$s, %1$s] FALLBACK_BASE_QUALITY=20 IGNORE_DUPLICATES=true USE_JDK_DEFLATER=false USE_JDK_INFLATER=false
  [Fri Aug 23 09:38:25 CEST 2024] Executing as kalexiou@kalexiou-IRTA on Linux 6.5.0-44-generic amd64; OpenJDK 64-Bit Server VM 22.0.1-internal-adhoc.conda.src; Deflater: Intel; Inflater: Intel; Provider GCS is not available; Picard version: 2.13.2-gridss
  INFO	2024-08-23 09:38:25	ExternalProcessFastqAligner	Invoking external aligner
  INFO	2024-08-23 09:38:25	ExternalProcessFastqAligner	bwa mem -K 10000000 -L 0,0 -t 4 genome.fa ./test.bam.gridss.working/test.bam.realign.0.fq
  mem: invalid option -- 'K'
  ERROR	2024-08-23 09:38:25	ExternalProcessHelper	Subprocess terminated with with exit status 1. Failed executing "bwa mem -K 10000000 -L 0,0 -t 4 genome.fa ./test.bam.gridss.working/test.bam.realign.0.fq". Can you run the alignment command from the command line? Is the aligner on PATH? Did you build an index with prefix genome.fa?
  [Fri Aug 23 09:38:25 CEST 2024] gridss.SoftClipsToSplitReads done. Elapsed time: 0.00 minutes.
  Runtime.totalMemory()=285212672
  ERROR	2024-08-23 09:38:25	CallVariants	Fatal exception thrown by worker thread.
  java.lang.RuntimeException: Subprocess terminated with with exit status 1. Failed executing "bwa mem -K 10000000 -L 0,0 -t 4 genome.fa ./test.bam.gridss.working/test.bam.realign.0.fq". Can you run the alignment command from the command line? Is the aligner on PATH? Did you build an index with prefix genome.fa?

It looks like that bwa mem doesn't recognise the -K argument. When I check bwa mem arguments (bwa-0.7.10-r789) I can see that indeed -K argument is absent, altough it is present in previous bwa versions. In issue #629, the user is running the same version of gridss as I do (gridss-2.13.2) and, though it reports a different kid of error, I can see that it passes through the bwa memcommand with success. But I cannot see which is the bwa version used.

Any help will be appreciated.

Best,
Kostas